We tested the cardiomyogenic potential of the human umbilical cord blood-derived mesenchymal stem cells (UCBMSCs). Both the number and function of stem cells may be depressed in senile patients with severe coronary risk factors. Therefore, stem cells obtained from such patients may not function well. For this reason, UCBMSCs are potentially a new cell source for stem cell-based therapy, since such cells can be obtained from younger populations and are being routinely utilized for clinical patients. The human UCBMSCs (5 × 103 per cm2) were cocultured with fetal murine cardiomyocytes ([CM] 1 × 105 per cm2). On day 5 of cocultivation, approximately half of the green fluorescent protein (GFP)-labeled UCBMSCs contracted rhythmically and synchronously, suggesting the presence of electrical communication between the UCBMSCs. The fractional shortening of the contracted UCBMSCs was 6.5% ± 0.7% (n = 20). The UCBMSC-derived cardiomyocytes stained positive for cardiac troponin-I (clear striation +) and connexin 43 (diffuse dot-like staining at the margin of the cell) by the immunocytochemical method. Cardiac troponin-I positive cardiomyocytes accounted for 45% ± 3% of GFP-labeled UCBMSCs. The cardiomyocyte-specific long action potential duration (186 ± 12 milliseconds) was recorded with a glass microelectrode from the GFP-labeled UCBMSCs. CM were observed in UCBMSCs, which were cocultivated in the same dish with mouse cardiomyocytes separated by a collagen membrane. Cell fusion, therefore, was not a major cause of CM in the UCBMSCs. Approximately half of the human UCBMSCs were successfully transdifferentiated into cardiomyocytes in vitro. UCBMSCs can be a promising cellular source for cardiac stem cell-based therapy.
Disclosure of potential conflicts of interest is found at the end of this article.
Autologous stem cells are believed to be a potential cellular source for stem cell-based therapy, since they have the ability to proliferate and differentiate into cardiomyocytes [1, , –4]. Many types of cells, such as embryonic stem cells [5, 6], myoblasts [7, 8], bone marrow hematopoietic cells [9, 10], and mesenchymal stem cells (MSCs) [11, –13], have been transplanted to restore damaged heart function in animal models. Autologous mononuclear cells [14, , –17] and myoblasts  have been injected into ischemic hearts clinically to improve impaired cardiac function. Despite the dramatic improvement of cardiac function by the stem-cell-based therapy in the animal model [10, 19], only modest effects were observed in the clinical study [14, , –17, 20]. One of the reasons for this may have been the extremely low rate of cardiomyogenesis of the stem cells in vitro and in vivo [2, 13, 21]. Therefore, the improvement of cardiac function may have been due to grafted stem cell-induced neovascularization [13, 22] and/or the paracrine effect . Another reason may have been the ages and disease histories of the patients. Recent papers have shown that the number and function of the circulating stem cells were depressed in older patients and in patients with diabetes mellitus [24, 25], suggesting that stem cells obtained from patients with coronary risk factors may not function well. This suggests limits to the utilization of autologous stem cells for the ischemic cardiomyopathy patient. On the other hand, in order to do allogenic stem cell transplantation, human leukocyte antigen (HLA)-type matching is very important for the stable survival of grafts. Therefore, the sample, which can be noninvasively collected from many volunteers, is a desirable source of stem cells due to the ease of establishing cell banks that can store all HLA-types.
Recently, umbilical cord blood (UCB) banking for transplantation of hematopoietic stem cells has become popular . If we can utilize UCB for heart failure patients, we can utilize this UCB stem cell bank network system immediately. UCB-derived stem cells may be superior to marrow-derived stem cells because they are obtained from infants. UCB contains circulating stem/progenitor cells, and the cells contained in UCB are known to be quite distinct from those contained in bone marrow and adult peripheral blood . UCB transplantation has been reported to improve cardiac function [28, –30]. That study, however, used a fraction of hematopoietic lineage and failed to show any clear evidence for cardiomyogenesis in vivo. In the present study, we focus on the mesenchymal lineage of UCB.
Isolation, characterization, and differentiation of clonally expanded umbilical cord blood-derived mesenchymal stem cells (UCBMSCs) have been reported [31, 32]. UCBMSCs have been found to have multipotency, and the immunophenotype of the clonally expanded cells is consistent with that reported for bone marrow mesenchymal stem cells [33, 34]. Kim et al.  showed modest but significant functional recovery of impaired cardiac function by transplantation of human unrestricted somatic stem cells obtained from umbilical cord blood that expressed mesenchymal cell surface markers ; therefore, mesenchymal lineage of the cells obtained from UCB may have potential therapeutic advantage in cardiac stem cell therapy. However, in vitro  and in vivo [34, 35], cardiomyogenic transdifferentiation ability have not yet been extensively studied. In the present study, we find that UCBMSCs have a strong potential for cardiomyogenic transdifferentiation.
Materials and Methods
Isolation and Cell Culture of UCBMSCs
The detailed isolation method has been described previously . A few colonies were found in the culture dish bottom 1 month after the collected cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS). One colony was trypsinized using a colony cylinder and then used for the experiment. We designated the monoclonal cell line as UCBMSC. The cells were prepared for infection with recombinant retroviruses expressing the human telomerase reverse transcriptase (TERT), as described previously [2, 33]. Stably transduced cells with an expanded life span were designated UCBMSC-TERT. The cells were cultured for further experiments under the approval of the Ethics Committee of our institute.
Preparation of Murine Fetal Cardiomyocytes
Fetal cardiomyocytes were obtained from the hearts of day 17 mouse fetuses . Hearts were minced with scissors and washed with phosphate-buffered saline (PBS), and the minced hearts were incubated in PBS with 0.05% trypsin and 0.25 mM EDTA (ethylenediamine-N,N,N′,N′-tetraacetic acid) (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) for 10 minutes at 37°C. After DMEM supplemented with 10% FBS was added, the cardiomyocytes were centrifuged at 1,000 rpm for 5 minutes. The pellet was then resuspended in 10 ml of DMEM with 10% FBS and incubated on glass dishes for 1 hour to separate the cardiomyocytes from fibroblasts. The floating cardiomyocytes were collected and replated at 1 × 105 per cm2.
Coculture System of UCBMSCs/UCBMSCs-TERT and Murine Fetal Cardiomyocytes
We employed a coculture system with fetal cardiomyocytes to induce cardiac transdifferentiation, since in vitro simulation of the heart by the environment has been shown to be an efficient means of inducing the transdifferentiation of human marrow-derived MSC . Cryopreserved UCBMSCs and UCBMSCs-TERT were used for the experiment. After thawing, the cells were cultured for at least two passages to stabilize the condition of the cell before the cardiomyogenic induction. UCBMSCs and UCBMSCs-TERT were labeled with enhanced green fluorescent protein (GFP) by recombinant adenovirus transfection as described previously . These cells were then exposed to 3 μM 5-azacytidine (5-azaC; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) for 24 hours to induce cell transdifferentiation or were left untreated. Then, 5 × 103 per cm2 of the cells were plated on the murine cardiomyocyte. The images were stored using a digital video system. The cell contraction was analyzed using a homemade image edge detection program made using Igor Pro 4 (WaveMetrics Inc., Portland, OR, http://www.wavemetrics.com). We administered 10 μM caffeine, 10 μM verapamil, or 1 μM thapsigargin to observe contraction of differentiated UCBMSCs.
A laser confocal microscope (FV1000; Olympus, Tokyo, http://www.olympus-global.com) was used for immunocytochemical analysis. The UCBMSCs and UCBMSCs-TERT were stained with mouse monoclonal anti-human cardiac troponin-I antibody (number 4T21 Lot 98/10-T21-C2; HyTest, Turku, Finland, http://www.hytest.fi) diluted 1:300, monoclonal anti-α actinin antibody (Sigma) diluted 1:300, or anti-connexin 43 antibody (Sigma) diluted 1:300. Nuclei were stained with 4′-6-diamidino-2-phenylindole (Wako Chemical, Osaka, Japan, http://www.wako-chem.co.jp/english) at 1:300. tetramethylrhodamine iso-thiocyanate (TRITC)-conjugated goat anti-mouse IgG (Sigma), TRITC-conjugated goat anti-rabbit IgG (Sigma), and Cy5-conjugated goat anti-mouse IgG (Chemicon, Temecula, CA, http://www.chemicon.com) were used as secondary antibodies.
Calculation of Induction Rate
After 1 week, UCBMSCs and UCBMSCs-TERT cultivated with or without murine fetal cardiomyocytes were detached from the dish by 0.1% trypsin and 0.25 mM EDTA for 5 minutes. The mass of cells obtained was then dissociated by 0.5% collagenase type-II (Worthington Biochemical, Lakewood, NJ, http://www.worthington-biochem.com) and 10 mM 2,3-butanedione monoxime (Sigma)-containing culture medium for 20–60 minutes. The isolated cells were seeded on poly-l-lysine coated dishes and stained. A confocal laser microscope was used to examine the cells. The cardiomyogenic induction rate was calculated as the fraction of human cardiac troponin-I-positive cells in the GFP-positive cells. The rate was calculated as the average from more than 10 separate experiments.
Examination of Chromosomes of UCBMSCs and Murine Cell Chimeras
Chromosomes from UCBMSCs cocultivated with murine cardiomyocyte for 1 week were stained by using a human chromosome-specific probe and a mouse chromosome-specific probe (Chromosome Science Labo, Hokkaido, Japan) and spectral karyotyping with fluorescence in situ hybridization (FISH) chromosome painting technique (Spectral Imaging, Vista, CA, http://www.spectral-imaging.com), according to the manufacturer's protocol.
Coculture of UCBMSCs-TERT and Murine Fetal Cardiomyocytes Separated by a Collagen Membrane
UCBMSCs-TERT and murine fetal cardiomyocytes were cocultured separately within the same dish. The murine fetal cardiomyocytes were seeded on top of a floating collagen film (CM-6; Koken, Tokyo, http://www.kokenmpc.co.jp/english), and the UCBMSCs-TERT were seeded on the bottom of the film. These two types of cells were, therefore, separated by a high-density atelocollagen film with a thickness of 30–40 μm, as shown in Figure 5E, that is permeable only for small molecules, less than 5,000 molecular weight (MW). After 1 week of cocultivation, the cells were analyzed immunocytochemically.
RNA Extraction and Reverse Transcriptase-Polymerase Chain Reaction
Total RNA was extracted from the UCBMSCs and UCBMSCs-TERT with RNeasy (Qiagen, Hilden, Germany, http://www1.qiagen.com). Human cardiac RNA was purchased (Clontech, Palo Alto, CA, http://www.clontech.com). RNA for reverse transcription-polymerase chain reaction (RT-PCR) was converted to cDNA with a first-strand cDNA synthesis kit (GE Healthcare, Buckinghamshire, U.K., http://www.gehealthcare.com) according to the manufacturer's recommendations. RT-PCR was performed by using primers for the following genes: Csx/Nkx-2.5, GATA4; cardiac hormones: human atrial natriuretic peptide (hANP), human brain natriuretic peptide (hBNP); cardiac structural proteins: cardiac troponin-I, cardiac troponin T, myosin heavy chain (MHC), myosin light chain-2a (MLC2a), cardiac-actin; ion channel: hyperpolarization-activated cyclic nucleotide-gated potassium channel 2 (HCN2); and 18s rRNA (18s rRNA was used as an internal control). PCR primers were prepared such that they would amplify the human but not the mouse genes  (supplemental online Table 1).
Action potentials (APs) from the spontaneously beating GFP-positive UCBMSCs and UCBMSCs-TERT were recorded by use of standard microelectrodes, as described previously . After the APs of the targeted cells were recorded, the dye (Alexa 568) was injected by electrophoresis (−5 nA for 10–20 seconds) to confirm the recorded APs obtained from GFP-positive cells. The extent of dye transfer was monitored under a fluorescent microscope.
Cardiomyogenic Transdifferentiation of UCBMSCs and UCBMSCs-TERT
On day 3 after starting the cocultivation, a few GFP-positive UCBMSCs and UCBMSCs-TERT started to contract (n = 68). On day 7, the beating of the murine cardiomyocytes stopped, whereas approximately half of the GFP-positive UCBMSCs and UCBMSCs-TERT beat strongly in a synchronized manner.
Immunocytochemistry revealed that a significant number of UCBMSCs and UCBMSCs-TERT expressing GFP were stained positive by the anti-human cardiac troponin-I antibody (Fig. 1A–1E, supplemental online Fig. 1A–1H). A clear striation pattern of cardiac troponin-I staining of UCBMSCs can be observed in higher magnification view (Fig. 1). Interestingly, troponin-I staining and GFP were observed alternately in a striated manner, suggesting that the troponin-I was expressed in the GFP-positive cells (Fig. 1 E). Clear striations were observed with red fluorescence of α-actinin in the differentiated UCBMSCs and UCBMSCs-TERT (Fig. 2B, 2H). Arrays of cardiomyocytes can be frequently observed (Fig. 2H). Connexin 43 staining (Fig. 2C, 2I) showed a clear and diffuse pattern around the margin of each GFP-positive cardiomyocyte, suggesting that these human transdifferentiated cardiomyocytes have tight electrical coupling with each other.
We also calculated the percentage of the human cardiac troponin-I-positive cells to determine the cardiomyogenic transdifferentiation rate of UCBMSCs and UCBMSCs-TERT. Since there was no essential difference between the UCBMSCs and UCBMSCs-TERT, calculated data from both cell types are averaged and shown in Figure 3. Although UCBMSCs without cocultivation did not show any troponin-I expression (Fig. 3H, 3K), 45% ± 3% of UCBMSCs became positive for human cardiac troponin-I antibody as a result of the cocultivation (Fig. 3B, 3E). It is noted that cardiomyogenic transdifferentiation could be observed in the cocultivated UCBMSCs and UCBMSCs-TERT without any 5-azaC pretreatment.
Cell fusion has been shown to be quite a rare phenomenon [4, 36]; however, it may contribute to the generation of cardiomyocytes in our system. Nuclear fusion between the cocultivated UCBMSCs-TERT and fetal murine cardiomyocytes was observed in only approximately 0.09% (2/2165) of the cocultivated cells by FISH analysis (Fig. 4A–4D). In the differentiated cardiomyocyte, there is no cell having double nuclei in the isolated GFP-positive UCBMSCs. Furthermore, in cocultures of UCBMSCs-TERT with fetal murine cardiomyocytes separated by a collagen membrane (Fig. 4E), we observed beating GFP-positive cells and human cardiac troponin-I expression (Fig. 4F–4L) (n = 8). Because these two cell types were not attached directly to each other, it was concluded that the cardiomyogenesis in the present study was mainly caused by the transdifferentiation of the UCBMSCs.
Expression of Cardiomyocyte-Specific Genes and Surface Markers of UCBMSCs and UCBMSCs-TERT
We analyzed the cocultured UCBMSCs and UCBMSCs-TERT in terms of gene expression and by immunocytochemistry and electrical recording. RT-PCR was performed with primers that hybridized with human cardiomyocyte-specific genes but not with the murine orthologues (second column from the right, Fig. 5A). Differentiated UCBMSCs-TERT expressed Csx/Nkx-2.5, GATA4, hANP, hBNP, cardiac-actin, MHC, MLC2a, cardiac troponin T, cardiac troponin-I, and HCN2. Interestingly, all of the analyzed genes except for the MHC and MLC2a were expressed in UCBMSCs and UCBMSCs-TERT before the induction, implying that UCBMSCs may have cardiomyogenic potential as a default state, like CMG cells, in which Csx/Nkx-2.5 and GATA4 are constitutively expressed before induction . Sequence analysis revealed that the sequences of the cDNAs matched those of the human genes.
Surface markers of the UCBMSCs-TERT were evaluated by flow cytometric analysis. The results showed that all of the UCBMSCs-TERT were positive for CD29 (integrin β1), CD44 (Pgp-1/ly-24), CD54, CD55, CD59, CDw90, CD105, CD157, CD164, CD166, and SSEA-4 and negative for CD14 (a marker for macrophage and dendritic cells), CD31 (platelet endothelial cell adhesion molecule-1), CD34, CD45 (leukocyte common antigen), CD117 (c-kit), CD133, CD140a, FIk-1, SSEA-1, and SSEA-3 (Fig. 5B). Our UCBMSCs are negative for CD34, CD45, Flk-1, and CD133, thus differing from hematopoietic stem cell and from circulating endothelial progenitor cells. It is noted that our UCBMSCs are weakly positive for SSEA4 , an embryonic stem cell marker. Thus, UCBMSCs may be more plastic for transdifferentiation than other somatic stem cells.
Functional Analysis of Differentiated UCBMSCs and UCBMSCs-TERT In Vitro
APs were recorded from spontaneously beating GFP-positive UCBMSCs and UCBMSCs-TERT. Alexa 568 was injected into cells via a recording microelectrode to stain the cells and confirm that the APs were generated by GFP-positive UCBMSCs (Fig. 6A, 6C). Since the dye did not diffuse into the murine cardiomyocytes, there were no tight cell-to-cell heterologous connections, that is, gap junctions. In most experiments, Alexa 568 diffused into the GFP-positive adjacent UCBMSCs and UCBMSCs-TERT, suggesting that homologous cell-to-cell connections had been established within 1 week after the start of cocultivation. The APs obtained from UCBMSCs and UCBMSCs-TERT showed clear cardiomyocyte-specific sustained plateaus. It was, therefore, concluded that they were the APs of cardiomyocytes, not of smooth muscle, nerve cells, or skeletal muscle (Fig. 6B, 6D). The measured parameters of the recorded APs were averaged (Fig. 6E). UCBMSCs and UCBMSCs-TERT did not have a marked pacemaker potential and had the character of working cardiomyocytes or ordinary cardiomyocytes. The rhythm of almost all of the UCBMSCs and UCBMSCs-TERT had become regular at 1 week. The fractional shortening (% FS) of the UCBMSCs and UCBMSCs-TERT was analyzed (Fig. 6F–6I) using a cell edge detection program. The GFP-positive cells contracted simultaneously within the whole visual field, suggesting tight electrical communication. There was no difference of % FS between the UCBMSCs and UCBMSCs-TERT. The % FS was augmented significantly by the administration of caffeine and inhibited by the administration of verapamil or thapsigargin (Fig. 7).
Physiologically Functioning Cardiomyocytes Can Be Generated from UCBMSCs In Vitro
Compared with the cardiomyogenic differentiation efficiency of the marrow-derived MSC (0.3%) , a significant number of the UCBMSCs transdifferentiated into cardiomyocytes in vitro in the present study. Generated cardiomyocytes showed physiologically functioning ability, that is, cardiomyocyte-specific APs with long duration (more than 100 milliseconds) and spontaneous contraction. The fact that each UCBMSC beats in a synchronized manner and the fact of the diffuse connexin 43 staining together suggest the formation of tight electrical coupling among the UCBMSCs. In our previous paper, we used the cells immediately after being quickly thawed from cryopreserved UCBMSCs, then failed to observe cardiomyogenic transdifferentiation in the small number of observations . Recently, however, we felt at least two passages should be required to stabilize and regain cardiomyogenic transdifferentiation ability in UCBMSCs and our several cell lines.
Highly Cardiomyogenic Differentiation Potential
In the marrow-derived stem cell, mesenchymal lineage has a cardiomyogenic transdifferentiation potential [2, 3]; hematopoietic cell lineage does not . This implies that mesenchymal lineage of the cell in UCB might have the ability to transdifferentiate into cardiomyocytes. Several in vivo experiments using UCB have shown feasible effects in restoring cardiac function in the myocardial infarction model [28, –30]. However, these experiments used CD34+ or CD133+ hematopoietic lineage of the cell in UCB and failed to show any clear evidence of cardiomyogenesis. Surface marker analysis revealed UCBMSCs as differing from hematopoietic stem cells and from circulating endothelial progenitor cells. Kögler et al.  reported that stem cells obtained from UCB, so-called unrestricted somatic stem cells (USSCs), have a pluripotent differential potential with a similar surface marker pattern, that is, negative for CD34 and CD45 and positive for CD29 and CD44, that is typical for mesenchymal cells. Furthermore, Kim et al.  showed that USSCs improved impaired cardiac function in vivo. Although the two papers showed modest evidence for cardiomyogenic potential of USSCs in vivo, experiments had not been extensively done to show the evidence of cardiac transdifferentiation. Finally, these papers failed to show clear evidence for cell fusion-independent cardiomyogenesis and efficiency of cardiomyogenic differentiation. In the present study, we show significant potential of cell fusion-independent cardiomyogenesis of UCBMSCs.
Comparisons with Other Stem Cells for Cardiology
Cardiac precursor cells (CPCs)  should be a promising stem cell source for cardiac regeneration therapy. However, CPCs failed to differentiate to the physiologically functioning cardiomyocyte in vitro, and cardiomyogenic differentiation efficiency in vivo was 29%–40%. Thus, cardiomyogenic differentiation efficiency might not be so markedly high compared with the UCBMSCs. Moreover, it is very difficult to match the donor-recipient HLA-type, and there is still a longstanding ethical problem. An embryonic stem cell is a pluripotential stem cell that has a cardiomyogenic differentiation potential. But there are still critical underlying problems, that is, teratoma formation , genomic alteration in long-term culture , and the ethical problem. Differing from embryonic stem cells, our RT-PCR data suggest constitutive expression of Nkx2.5/Csx and GATA4 and other cardiac structure mRNA with the ability of self-renewal. This suggests that some population of the UCBMSCs has cardiomyogenic potential as the default state, and they may be termed cardiac precursor cells in light of their biological features. Recently, we reported that human endometrial gland-derived mesenchymal cells also have a high cardiomyogenic potential . This suggests that they may be a stem cell source for heart disease. However, for male patients, there is no choice for autologous transplantation of this cell and no running stem cell bank for this cell. On the other hand, if UCBMSCs were isolated and frozen at the time of birth, they could later be thawed for use by the donor who required cardiac stem cell therapy at a later age. Furthermore, UCB banking has played a major role for hematopoietic stem cell transplantation for leukemia treatment. If we utilize a world-wide UCB bank system for cardiac stem cell therapy, we may be able to utilize UCBMSCs for cardiac stem cell therapy in the near future. Since several reports showed that mesenchymal cells cause immunological tolerance [42, –44], we speculate that only a minimum administration of immunosuppressive agents may be sufficient to control rejection of the allogenic UCBMSC transplantation, if we match the other MHC antigen by utilizing the stem cell bank system.
From a single stem cell we can obtain approximately 232 cells with extremely high cardiomyogenic potential; however, the number of MSCs in the UCB is quite low, as was described previously [33, 34]. Thus, further experiments should be done to establish a method to collect the UCBMSCs efficiently. The transfection of the TERT gene may alter the phenotype of UCBMSCs to some extent. However, TERT-gene transfection was not essential for causing cardiomyogenic differentiation of UCBMSCs, and there was no essential difference between the UCBMSCs and UCBMSCs-TERT in the present study.
Our in vitro cardiomyogenic induction system provided a substantial environmental factor to cause cardiomyogenic transdifferentiation of UCBMSCs in vitro; however, specific key factors (e.g., humoral factors) for cardiomyogenesis were still unclear. It is still undetermined whether such key factors for cardiomyogenesis are sufficiently provided by the surrounding host heart when UCBMSCs are engrafted in vivo. We believe that the definition of these specific factors in vitro should be extremely important to improve cardiomyogenesis in situ; therefore, in the present study, we focused on in vitro cardiomyogenesis of UCBMSCs.
Cell fusion is a rare phenomenon (0.6%–0.05%) , and the frequency of nuclear fusion was low (0.1%) in the present study. On the other hand, the cardiomyogenic differentiation efficiency of UCBMSCs was extremely high (44.9% ± 3.6%). Furthermore, a 40-μm-thick atelocollagen membrane is not permeable for molecules larger than 5,000 MW, and no cell migration from the top of the membrane to the bottom was observed in our culture condition. On this basis, we concluded that cell fusion did not play a major role in the UCBMSC-derived cardiomyogenesis in the present study.
Our major findings in the present study are: (a) for the first time, physiologically functioning cardiomyocytes were transdifferentiated from human UCBMSCs in vitro; (b) the observed cardiomyogenic transdifferentiation, independent of cell fusion, was approximately 44.9% ± 3.6% of UCBMSCs; and (c) cocultivation with fetal murine cardiomyocytes alone without other transdifferentiation factors, that is, 5-azaC, is sufficient for cardiomyogenesis in our system. Therefore, UCBMSCs may be a promising cellular source for cardiac stem cell-based therapy, by which cardiomyogenesis can be expected.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.